Share

The Economics of Nuclear Power

(Updated September 2014)

  • Nuclear power is cost competitive with other forms of electricity generation, except where there is direct access to low-cost fossil fuels.
  • Fuel costs for nuclear plants are a minor proportion of total generating costs, though capital costs are greater than those for coal-fired plants and much greater than those for gas-fired plants.
  • Providing incentives for long-term, high-capital investment in deregulated markets where short-term price signals present a challenge in securing a diversified and reliable electricity supply system.
  • In assessing the economics of nuclear power, decommissioning and waste disposal costs are fully taken into account.

Assessing the relative costs of new generating plants utilising different technologies is a complex matter and the results depend crucially on location. Coal is, and will probably remain, economically attractive in countries such as China, the USA and Australia with abundant and accessible domestic coal resources as long as carbon emissions are cost-free. Gas is also competitive for base-load power in many places, particularly using combined-cycle plants, though rising gas prices have removed much of the advantage.

Nuclear power plants are expensive to build but relatively cheap to run. In many places, nuclear energy is competitive with fossil fuels as a means of electricity generation. Waste disposal and decommissioning costs are included in the operating costs. If the social, health and environmental costs of fossil fuels are also taken into account, the economics of nuclear power are outstanding.

See also the December 2005 World Nuclear Association report (pdf 310 kB) The New Economics of Nuclear Power.

Assessing the costs of nuclear power

The economics of nuclear power involves consideration of several aspects:

Capital costs, which include the cost of site preparation, construction, manufacture, commissioning and financing a nuclear power plant. Building a large-scale nuclear reactor takes thousands of workers, huge amounts of steel and concrete, thousands of components, and several systems to provide electricity, cooling, ventilation, information, control and communication. To compare different power generation technologies the capital costs must be expressed in terms of the generating capacity of the plant (for example as dollars per kilowatt). Capital costs may be calculated with the financing costs included or excluded. If financing costs are included then the capital costs change in proportion to the length of time it takes to build and commission the plant and with the interest rate or mode of financing employed. It is normally termed the ‘investment cost’. If the financing costs are excluded from the calculation the capital costs is called the ‘overnight cost’, because it imagines that the plant appeared fully built overnight.

Plant operating costs, which include the costs of fuel, operation and maintenance (O&M), and a provision for funding the costs of decommissioning the plant and treating and disposing of used fuel and wastes. Operating costs may be divided into ‘fixed costs’ that are incurred whether or not the plant is generating electricity and ‘variable costs’, which vary in relation to the output. Normally these costs are expressed relative to a unit of electricity (for example, cents per kilowatt-hour) to allow a consistent comparison with other energy technologies. To calculate the operating cost of a plant over its whole life (including the costs of decommissioning and used fuel and waste management), we must estimate the ‘levelised’ cost at present value. It represents the price that the electricity must fetch if the project is to break even (after taking account of the opportunity cost of capital through the application of a discount rate).

External costs to society from the operation, which in the case of a nuclear power is usually assumed to be zero, but could include the costs of dealing with a serious accident that are beyond the insurance limit and in practice need to be picked up by the government. The regulations that control nuclear power typically require the plant operator to make a provision for disposing of any waste, thus these costs are ‘internalised’ (and are not external). Electricity generation from fossil fuels is not regulated in the same way, and therefore the operators of such thermal power plants do not yet internalise the costs of greenhouse gas emission or of other gases and particulates released into the atmosphere. Including these external costs in the calculation gives nuclear power a significant advantage over fossil fuelled electricity generation.

Considering these costs in turn, with information from numerous studies:

Capital cost

Construction costs comprise several things: the bare plant cost (usually identified as engineering-procurement-construction – EPC – cost), the owner's costs (land, cooling infrastructure, administration and associated buildings, site works, switchyards, project management, licences, etc.), cost escalation and inflation. Owner's costs may include some transmission infrastructure. Recent studies have shown an increase in the capital cost of building both conventional and nuclear power plants.

The term "overnight capital cost" is often used, meaning EPC plus owners’ costs and excluding financing, escalation due to increased material and labour costs, and inflation. Construction cost – sometimes called 'all-in cost' – adds to overnight cost any escalation and interest during construction and up to the start of construction. It is expressed in the same units as overnight cost and is useful for identifying the total cost of construction and for determining the effects of construction delays. In general the construction costs of nuclear power plants are significantly higher than for coal- or gas-fired plants because of the need to use special materials, and to incorporate sophisticated safety features and back-up control equipment. These contribute much of the nuclear generation cost, but once the plant is built the cost variables are minor. The OECD Nuclear Energy Agency’s (NEA) calculation of the overnight cost for a nuclear power plant built in the OECD rose from about US$ 1,900/kWe at the end of the 1990s to US$ 3,850/kWe in 2009.

The NEA figures for the 1990s must be treated with caution as they are not in line with some other data sources. The US Energy Information Administration (EIA) calculated that, in constant 2002 values, the realized real overnight cost of a nuclear power plant built in the USA grew from US$ 1,500/kWe in the early 1960s to US$ 4,000/kWe in the mid-1970s. The EIA cited increased regulatory requirements (including design changes that required plants to be back-fitted with modified equipment), licensing problems, project management problems and mis-estimation of costs and demand as the factors contributing to the increase during the 1970s. Its 2010 report, Updated Capital Cost Estimates for Electricity Generation Plants, gave an estimate for a new nuclear plant of US$ 5,339/kW.

There is also significant variation of capital costs by country, particularly between the emerging industrial economies of East Asia and the mature markets of Europe and North America, which has a variety of explanations, including differential labour costs, more experience in the recent building of reactors, economies of scale from building multiple units and streamlined licensing and project management within large civil engineering projects. With few new orders, the data set for new build costs is lacking. The shift to Generation III reactors has added further uncertainty. Other non-nuclear generation technologies also show variation and as do major infrastructure projects such as roads and bridges, depending upon where they are built. However, the variation is particularly crucial for nuclear as its economics depend so much on minimising its capital investment cost.

The French national audit body, the Cour des comptes, said in 2012 that the overnight capital costs of building NPPs increased over time from € 1,070/kWe (at 2010 prices) when the first of the 50 PWRs was built at Fessenheim (commissioned in 1978) to € 2,060/kWe when Chooz 1 and 2 were built in 2000, and to a projected € 3,700/kWe for the Flamanville EPR. It can be argued that much of this escalation relates to the smaller magnitude of the programme by 2000 (compared with when the French were commissioning 4-6 new PWRs per year in the 1980s) and to the subsequent loss of economies of scale.

In several countries, notably the UK, there is a trend to greater vendor involvement in financing projects, but with an intention to relinquish equity once the plant is running.

A presentation by Dr. N. Barkatullah, UAE Regulation & Supervision Bureau, at the World Nuclear Association 2014 Symposium showed the risk in construction costs (per kilowatt of capacity), much of it due to financing cost incurred by delays:

Challenge: Construction Risk

The same presentation showed the following ranges of figures for overnight capital cost in different parts of the world:

Challenge: NPP investment cost uncertainty

By way of contrast, China has stated that it expects its costs for plants under construction to come in at less than $2000/kW and that subsequent units should be in the range of $1600/kW. This estimate is for the AP1000 design, the same as used by EIA for the USA. This would mean that an AP1000 in the USA would cost about three times as much as the same plant built in China. Different labour rates in the two countries are only part of the explanation. Standardised design, numerous units being built, and increased localisation are all significant factors in China.

Financing costs will depend on the rate of interest on debt, the debt-equity ratio, and if it is regulated, how the capital costs are recovered. There must also be an allowance for a rate of return on equity, which is risk capital.

Long construction periods will push up financing costs, and in the past they have done so spectacularly. In Asia construction times have tended to be shorter, for instance the new-generation 1300 MWe Japanese reactors which began operating in 1996 and 1997 were built in a little over four years, and 48 to 54 months is typical projection for plants today. See also Construction Risk graphic above.

Operating costs

Fuel costs have from the outset given nuclear energy an advantage compared with coal, oil and gas-fired plants. Uranium, however, has to be processed, enriched and fabricated into fuel elements, and about half of the cost is due to enrichment and fabrication. In the assessment of the economics of nuclear power allowances must also be made for the management of radioactive used fuel and the ultimate disposal of this used fuel or the wastes separated from it. But even with these included, the total fuel costs of a nuclear power plant in the OECD are typically about a third of those for a coal-fired plant and between a quarter and a fifth of those for a gas combined-cycle plant. The US Nuclear Energy Institute suggests that for a coal-fired plant 78% of the cost is the fuel, for a gas-fired plant the figure is 89%, and for nuclear the uranium is about 14%, or double that to include all front end costs.

In June 2013, the approx. US $ cost to get 1 kg of uranium as UO2 reactor fuel (at current spot uranium price):

Uranium: 8.9 kg U3O8 x $130 US$ 1160
Conversion: 7.5 kg U x $11 US$ 83
Enrichment: 7.3 SWU x $120 US$ 880
Fuel fabrication: per kg US$ 240
Total, approx:   US$ 2360

At 45,000 MWd/t burn-up this gives 360,000 kWh electrical per kg, hence fuel cost: 0.66 c/kWh.

Fuel costs are one area of steadily increasing efficiency and cost reduction. For instance, in Spain the nuclear electricity cost was reduced by 29% over 1995-2001. This involved boosting enrichment levels and burn-up to achieve 40% fuel cost reduction. Prospectively, a further 8% increase in burn-up will give another 5% reduction in fuel cost.

Uranium has the advantage of being a highly concentrated source of energy which is easily and cheaply transportable. The quantities needed are very much less than for coal or oil. One kilogram of natural uranium will yield about 20,000 times as much energy as the same amount of coal. It is therefore intrinsically a very portable and tradable commodity.

The contribution of fuel to the overall cost of the electricity produced is relatively small, so even a large fuel price escalation will have relatively little effect (see below). Uranium is abundant and widely available.

There are other possible savings. For example, if used fuel is reprocessed and the recovered plutonium and uranium is used in mixed oxide (MOX) fuel, more energy can be extracted. The costs of achieving this are large, but are offset by MOX fuel not needing enrichment and particularly by the smaller amount of high-level wastes produced at the end. Seven UO2 fuel assemblies give rise to one MOX assembly plus some vitrified high-level waste, resulting in only about 35% of the volume, mass and cost of disposal. 

Operating costs include operating and maintenance (O&M) plus fuel. Fuel cost figures include used fuel management and final waste disposal. These costs, while usually external for other technologies, are internal for nuclear power (i.e. they have to be paid or set aside securely by the utility generating the power, and the cost passed on to the customer in the actual tariff).

This 'back end' of the fuel cycle, including used fuel storage or disposal in a waste repository, contributes up to 10% of the overall costs per kWh – rather less if there is direct disposal of used fuel rather than reprocessing. The $26 billion US used fuel program is funded by a 0.1 cent/kWh levy.

Decommissioning costs are about 9-15% of the initial capital cost of a nuclear power plant. But when discounted, they contribute only a few percent to the investment cost and even less to the generation cost. In the USA they account for 0.1-0.2 cent/kWh, which is no more than 5% of the cost of the electricity produced. 

System costs

System costs are the total costs above plant-level costs (capital and operating) to supply electricity at a given load and given level of security of supply. They include grid connection, extension and reinforcement, short-term balancing costs and long-term costs of maintaining adequate back-up.

They are external to the building and operation of any power plant, but must be paid by the electricity consumer, usually as part of the transmission and distribution cost. From a government policy point of view they are just as significant as the actual generation cost, but are seldom factored in to comparisons among different supply options, especially comparing base-load with dispersed renewables. In fact that the total system cost should be analysed when introducing new power generating capacity on the grid. Any new power plant likely requires changes to the grid, and hence incurs a significant cost for power supply that must be accounted for. But this cost for large base-load plants is small compared with integrating renewables to the grid.

The integration of intermittent renewable supply on a preferential basis despite higher unit cost creates significant diseconomies for dispatchable supply, as is now becoming evident in Germany, Austria and Spain, compromising security of supply and escalating costs.

An OECD study found that the integration of large shares of intermittent renewable electricity is a major challenge for the electricity systems of OECD countries and for dispatchable generators such as nuclear. Grid-level system costs for variable renewables are large ($15-80/MWh) but depend on country, context and technology (onshore wind < offshore wind < Solar PV). Nuclear system cost is $1-3/MWh.

See also paper on Electricity Transmission Grids

External costs

External costs are not included in the building and operation of any power plant, and are not paid by the electricity consumer, but by the community generally. The external costs are defined as those actually incurred in relation to health and the environment, and which are quantifiable but not built into the cost of the electricity.

The report of a major European study of the external costs of various fuel cycles, focusing on coal and nuclear, was released in mid 2001 – ExternE. It shows that in clear cash terms nuclear energy incurs about one tenth of the costs of coal. If these costs were in fact included, the EU price of electricity from coal would double and that from gas would increase 30%. These are without attempting to include the external costs of global warming.

The European Commission launched the project in 1991 in collaboration with the US Department of Energy, and it was the first research project of its kind "to put plausible financial figures against damage resulting from different forms of electricity production for the entire EU". The methodology considers emissions, dispersion and ultimate impact. With nuclear energy the risk of accidents is factored in along with high estimates of radiological impacts from mine tailings (waste management and decommissioning being already within the cost to the consumer). Nuclear energy averages 0.4 euro cents/kWh, much the same as hydro, coal is over 4.0 cents (4.1-7.3), gas ranges 1.3-2.3 cents and only wind shows up better than nuclear, at 0.1-0.2 cents/kWh average. NB these are the external costs only.

See also paper on Energy subsidies and external costs.

Comparing the economics of different forms of electricity generation

In 2013 the US Energy Information Administration published figures for the average levelized costs per unit of output for generating technologies to be brought on line in 2018, as modeled for its Annual Energy Outlook. These show advanced nuclear, natural gas (advanced combustion turbine), and conventional coal in the bracket 10-11c/kWh. Combined cycle natural gas is 6.6 cents, advanced coal with CCS 13.6 cents, and among the non-dispatchable technologies: wind onshore 8.7 cents, solar PV 14.4 cents, offshore wind 22.2 cents and solar thermal 26.2 c/kWh. The actual capital cost of nuclear is about the same as coal, and very much more than any gas option.

A 2010 OECD study Projected Costs of generating Electricity set out some actual costs of electricity generation, from which the following figures are taken:

Actual Costs of Electricity (US cents/kWh)

Technology region or country At 10% discount rate At 5% discount rate
Nuclear OECD Europe 8.3-13.7 5.0-8.2
  China 4.4-5.5 3.0-3.6
Black coal with CCS OECD Europe 11.0 8.5
Brown coal with CCS OECD Europe 9.5-14.3 6.8-9.3
CCGT with CCS OECD Europe 11.8 9.8
Large hydro-electric OECD Europe 14.0-45.9 7.4-23.1
  China: 3 Gorges 5.2 2.9
  China: other 2.3-3.3 1.2-1.7
Onshore wind OECD Europe 12.2-23.0 9.0-14.6
  China 7.2-12.6 5.1-8.9
Offshore wind OECD Europe 18.7-26.1 13.8-18.8
Solar photovoltaic OECD Europe 38.8-61.6 28.7-41.0
  China 18.7-28.3 12.3-18.6

Source: OECD/IEA-NEA, 2010, Projected Costs of Generating Electricity, Tables 3.7
This shows the levelised cost, which is the average cost of producing electricity including capital, finance, owner's costs on site, fuel and operation over a plant's lifetime.

It is important to distinguish between the economics of nuclear plants already in operation and those at the planning stage. Once capital investment costs are effectively “sunk”, existing plants operate at very low costs and are effectively “cash machines”. Their operations and maintenance (O&M) and fuel costs (including used fuel management) are, along with hydropower plants, at the low end of the spectrum and make them very suitable as base-load power suppliers. This is irrespective of whether the investment costs are amortized or depreciated in corporate financial accounts – assuming the forward or marginal costs of operation are below the power price, the plant will operate.

US figures for 2012 published by NEI show the general picture, with nuclear generating power at 2.40 c/kWh, compared with coal at 3.27 cents and gas at 3.40 cents.

U.S. Electricity Prodiction Costs, 1995-2012
Click for high resolution graphic

Note: the above data refer to fuel plus operation and maintenance costs only, they exclude capital, since this varies greatly among utilities and states, as well as with the age of the plant.

A Finnish study in 2000 also quantified fuel price sensitivity to electricity costs:

The impact of fuel costs on electricity generation costs

These show that a doubling of fuel prices would result in the electricity cost for nuclear rising about 9%, for coal rising 31% and for gas 66%. Gas prices have since risen significantly.

The impact of varying the uranium price in isolation is shown below in a worked example of a typical US plant, assuming no alteration in the tails assay at the enrichment plant.

Effect of U Price

Doubling the uranium price (say from $25 to $50 per lb U3O8) takes the fuel cost up from 0.50 to 0.62 US cents per kWh, an increase of one quarter, and the expected cost of generation of the best US plants from 1.3 US cents per kWh to 1.42 cents per kWh (an increase of almost 10%). So while there is some impact, it is comparatively minor, especially by comparison with the impact of gas prices on the economics of gas generating plants. In these, 90% of the marginal costs can be fuel. Only if uranium prices rise to above $100 per lb U3O8 ($260 /kgU) and stay there for a prolonged period (which seems very unlikely) will the impact on nuclear generating costs be considerable.

Nevertheless, for nuclear power plants operating in competitive power markets where it is impossible to pass on any fuel price increases (ie the utility is a price-taker), higher uranium prices will cut corporate profitability. Yet fuel costs have been relatively stable over time – the rise in the world uranium price between 2003 and 2007 added to generation costs, but conversion, enrichment and fuel fabrication costs did not followed the same trend.

In February 2014 the US Nuclear Energy Institute presented figures from the Electric Utility Cost Group on US generating costs comprising fuel, capital and operating costs for 61 nuclear sites in 2012. The average came to $44/MWh, being $50.54 for single-unit plants and $39.44 for multi-unit plants (all two-unit except Browns Ferry, Oconee and Palo Verde). The $44 represented a 58% increase in ten years, largely due to a threefold increase in capital expenditure on plants which were mostly old enough to be fully depreciated. Over half of the capital expenditure (51%) in 2012 related to power uprates and licence renewals, while 26% was for equipment replacement.  

For prospective new nuclear plants, the fuel component is even less significant (see below). The typical front end nuclear fuel cost is typically only 15-20% of the total, as opposed to 30-40% for operating nuclear plants.

Competitiveness in the context of increasing use of power from renewable sources, which are legally preferred, is a major issue today. The most important renewable sources are intermittent by nature, which means that their supply to the electricity system does not necessarily match demand from customers. In power grids where renewable sources of generation make a significant contribution, intermittency forces other generating sources to ramp up their supply or power down at short notice. This volatility can have a large impact on non-intermittent generators’ profitability.A variety of responses to the challenge of intermittent generation are possible. Two options currently being implemented are increased conventional plant flexibility and increased grid capacity and coverage. Flexibility is seen as most applicable to gas and coal fired generators, but nuclear reactors, normally regarded as base-load producers, also have the ability to load-follow, eg, by the use of ‘grey rods’ to modulate the reaction speed.

As the scale of intermittent generating capacity increases however, more significant measures will be required. The establishment and extension of capacity mechanisms, which offer payments to generators prepared to guarantee supply for defined periods, are now under serious consideration within the EU. Capacity mechanisms can in theory provide security of supply to desired levels but at a price which might be high, for example, Morgan Stanley has estimated that investors in a 800 MWe gas plant providing for intermittent generation would require payments of €80 million per year whilst Ecofys calculate that a 4 GWe reserve in Germany would cost €140-240/year. Almost by definition, investors in conventional plant designed to operate intermittently will face low and uncertain load factors and will therefore demand significant capacity payments in return for the investment decision. In practice, until the capacity mechanism has been reliably implemented, investors are likely to withhold investment. Challenges for EU power market integration are expected to result from differences between member state capacity mechanisms.

Future cost competitiveness

Understanding the cost of new generating capacity and its output requires careful analysis of what is in any set of figures. There are three broad components: capital, finance and operating costs. Capital and financing costs make up the project cost. 

Calculations of relative generating costs are made using levelised costs, meaning average costs of producing electricity including capital, finance, owner's costs on site, fuel and operation over a plant's lifetime, with provision for decommissioning and waste disposal.

It is important to note that capital cost figures quoted by reactor vendors, or which are general and not site-specific, will usually just be for EPC costs. This is because owner's costs will vary hugely, most of all according to whether a plant is Greenfield or at an established site, perhaps replacing an old plant.

There are several possible sources of variation which preclude confident comparison of overnight or EPC (Engineering, Procurement & Construction) capital costs – eg whether initial core load of fuel is included. Much more obvious is whether the price is for the nuclear island alone (Nuclear Steam Supply System) or the whole plant including turbines and generators – all the above figures include these. Further differences relate to site works such as cooling towers as well as land and permitting – usually they are all owner's costs as outlined earlier in this section. Financing costs are additional, adding typically around 30%, and finally there is the question of whether cost figures are in current (or specified year) dollar values or in those of the year in which spending occurs. 

Major studies on future cost competitiveness

There have been many studies carried out examining the economics of future generation options, and the following are merely the most important and also focus on the nuclear element.

The 2010 OECD study Projected Costs of generating Electricity compared 2009 data for generating base-load electricity by 2015 as well as costs of power from renewables, and showed that nuclear power was very competitive at $30 per tonne CO2 cost and low discount rate. The study comprised data for 190 power plants from 17 OECD countries as well as some data from Brazil, China, Russia and South Africa. It used levelised lifetime costs with carbon price internalised (OECD only) and discounted cash flow at 5% and 10%, as previously. The precise competitiveness of different base-load technologies depended very much on local circumstances and the costs of financing and fuels.

Nuclear overnight capital costs in OECD ranged from US$ 1556/kW for APR-1400 in South Korea through $3009 for ABWR in Japan, $3382/kW for Gen III+ in USA, $3860 for EPR at Flamanville in France to $5863/kW for EPR in Switzerland, with world median $4100/kW. Belgium, Netherlands, Czech Rep and Hungary were all over $5000/kW. In China overnight costs were $1748/kW for CPR-1000 and $2302/kW for AP1000, and in Russia $2933/kW for VVER-1150. EPRI (USA) gave $2970/kW for APWR or ABWR, Eurelectric gave $4724/kW for EPR. OECD black coal plants were costed at $807-2719/kW, those with carbon capture and compression (tabulated as CCS, but the cost not including storage) at $3223-5811/kW, brown coal $1802-3485, gas plants $635-1747/kW and onshore wind capacity $1821-3716/kW. (Overnight costs were defined here as EPC, owner's costs and contingency, but excluding interest during construction.)

OECD electricity generating cost projections for year 2010 on - 5% discount rate, c/kWh

country nuclear coal coal with CCS Gas CCGT Onshore wind
Belgium 6.1 8.2 - 9.0 9.6
Czech R 7.0 8.5-9.4 8.8-9.3 9.2 14.6
France 5.6 - - - 9.0
Germany 5.0 7.0-7.9 6.8-8.5 8.5 10.6
Hungary 8.2 - - - -
Japan 5.0 8.8 - 10.5 -
Korea 2.9-3.3 6.6-6.8 - 9.1 -
Netherlands 6.3 8.2 - 7.8 8.6
Slovakia 6.3 12.0 - - -
Switzerland 5.5-7.8 - - 9.4 16.3
USA 4.9 7.2-7.5 6.8 7.7 4.8
China* 3.0-3.6 5.5 - 4.9 5.1-8.9
Russia* 4.3 7.5 8.7 7.1 6.3
EPRI (USA) 4.8 7.2 - 7.9 6.2
Eurelectric 6.0 6.3-7.4 7.5 8.6 11.3

* For China and Russia: 2.5c is added to coal and 1.3c to gas as carbon emission cost to enable sensible comparison with other data in those fuel/technology categories, though within those countries coal and gas will in fact be cheaper than the Table above suggests.
Source: OECD/IEA NEA 2010, table 4.1.

At 5% discount rate comparative costs are as shown above. Nuclear is comfortably cheaper than coal and gas in all countries. At 10% discount rate (below) nuclear is still cheaper than coal in all but the Eurelectric estimate and three EU countries, but in these three gas becomes cheaper still. Coal with carbon capture is mostly more expensive than either nuclear or paying the $30 per tonne for CO2 emissions, though the report points out "great uncertainties" in the cost of projected CCS. Also, investment cost becomes a much greater proportion of power cost than with 5% discount rate.

OECD electricity generating cost projections for year 2010 on - 10% discount rate, c/kWh

country nuclear coal coal with CCS Gas CCGT Onshore wind
Belgium 10.9 10.0 - 9.3-9.9 13.6
Czech R 11.5 11.4-13.3 13.6-14.1 10.4 21.9
France 9.2 - - - 12.2
Germany 8.3 8.7-9.4 9.5-11.0 9.3 14.3
Hungary 12.2 - - - -
Japan 7.6 10.7 - 12.0 -
Korea 4.2-4.8 7.1-7.4 - 9.5 -
Netherlands 10.5 10.0 - 8.2 12.2
Slovakia 9.8 14.2 - - -
Switzerland 9.0-13.6 - - 10.5 23.4
USA 7.7 8.8-9.3 9.4 8.3 7.0
China* 4.4-5.5 5.8 - 5.2 7.2-12.6
Russia* 6.8 9.0 11.8 7.8 9.0
EPRI (USA) 7.3 8.8 - 8.3 9.1
Eurelectric 10.6 8.0-9.0 10.2 9.4 15.5

* For China and Russia: 2.5c is added to coal and 1.3c to gas as carbon emission cost to enable sensible comparison with other data in those fuel/technology categories, though within those countries coal and gas will in fact be cheaper than the Table above suggests.
Source: OECD/IEA NEA 2010, table 4.1.

A 2004 report on The Economic Future of Nuclear Power from from the University of Chicago, funded by the US Department of Energy, compared the levelised power costs of future nuclear, coal, and gas-fired power generation in the USA. Various nuclear options were covered, and for an initial ABWR or AP1000 they range from 4.3 to 5.0 c/kWh on the basis of overnight capital costs of $1200 to $1500/kW, 60 year plant life, 5 year construction and 90% capacity. Coal gives 3.5 - 4.1 c/kWh and gas (CCGT) 3.5 - 4.5 c/kWh, depending greatly on fuel price.

The levelised nuclear power cost figures include up to 29% of the overnight capital cost as interest, and the report notes that up to another 24% of the overnight capital cost needs to be added for the initial unit of a first-of-a-kind advanced design such as the AP1000, defining the high end of the range above. For more advanced plants such as the EPR or SWR1000, overnight capital cost of $1800/kW is assumed and power costs are projected beyond the range above. However, considering a series of eight units of the same kind and assuming increased efficiency due to experience which lowers overnight capital cost, the levelised power costs drop 20% from those quoted above and where first-of-a-kind engineering costs are amortised (eg the $1500/kW case above), they drop 32%, making them competitive at about 3.4 c/kWh.

Nuclear plant: projected electrcity costs (c/kWh)

Overnight capital cost $/kW 1200 1500 1800
First unit 7 yr build, 40 yr life
5.3
6.2
7.1
  5 yr build, 60 yr life
4.3
5.0
5.8
4th unit 7 yr build, 40 yr life
4.5
4.5
5.3
  5 yr build, 60 yr life *
3.7
3.7
4.3
8th unit 7 yr build, 40 yr life
4.2
4.2
4.9
  5 yr build, 60 yr life *
3.4
3.4
4.0

* calculated from above data

The study also shows that with a minimal carbon control cost impact of 1.5 c/kWh for coal and 1.0 c/kWh for gas superimposed on the above figures, nuclear is even more competitive. But more importantly it goes on to explore other policy options which would offset investment risks and compensate for first-of-a-kind engineering costs to encourage new nuclear investment, including investment tax breaks, and production tax credits phasing out after 8 years. (US wind energy gets a production tax credit which has risen to 2.1 c/kWh.)

In May 2009 an update of a heavily-referenced 2003 MIT study on The Future of Nuclear Power was published. This said that "since 2003 construction costs for all types of large-scale engineered projects have escalated dramatically. The estimated cost of constructing a nuclear power plant has increased at a rate of 15% per year heading into the current economic downturn. This is based both on the cost of actual builds in Japan and Korea and on the projected cost of new plants planned for in the United States. Capital costs for both coal and natural gas have increased as well, although not by as much. The cost of natural gas and coal that peaked sharply is now receding. Taken together, these escalating costs leave the situation [of relative costs] close to where it was in 2003." The overnight capital cost was given as $4000/kW, in 2007 dollars. Applying the same cost of capital to nuclear as to coal and gas, nuclear came out at 6.6 c/kWh, coal at 8.3 cents and gas at 7.4 cents, assuming a charge of $25/tonne CO2 on the latter.

The French Energy & Climate Directorate published in November 2008 an update of its earlier regular studies on relative electricity generating costs. This shied away from cash figures to a large extent due to rapid changes in both fuel and capital, but showed that at anything over 6000 hours production per year (68% capacity factor), nuclear was cheaper than coal or gas combined cycle (CCG). At 100% capacity CCG was 25% more expensive than nuclear. At less than 4700 hours per year CCG was cheapest, all without taking CO2 cost into account. 

With the nuclear plant fixed costs were almost 75% of the total, with CCG they were less than 25% including allowance for CO2 at $20/t. Other assumptions were 8% discount rate, gas at 6.85 $/GJ, coal at EUR 60/t. The reference nuclear unit is the EPR of 1630 MWe net, sited on the coast, assuming all development costs being borne by Flamanville 3, coming on line in 2020 and operating only 40 of its planned 60 years. Capital cost apparently EUR 2000/kW. Capacity factor 91%, fuel enrichment is 5%, burnup 60 GWd/t and used fuel is reprocessed with MOX recycle. In looking at overall fuel cost, uranium at $52/lb made up about 45% of it, and even though 3% discount rate was used for back-end the study confirmed the very low cost of waste in the total - about 13% of fuel cost, mostly for reprocessing.

At the end of 2008 EdF updated the overnight cost estimate for Flamanville 3 EPR (the first French EPR, but with some supply contracts locked in before escalation) to EUR 4 billion in 2008 Euros (EUR 2434/kW), and electricity cost 5.4 cents/kWh (compared with 6.8 c/kWh for CCGT and 7.0 c/kWh for coal, "with lowest assumptions" for CO2 cost). These costs were confirmed in mid 2009, when EdF had spent nearly EUR 2 billion. In July 2010 EdF revised the overnight cost to about EUR 5 billion.

A detailed study of energy economics in Finland published in mid 2000 was important in making the strong case for additional nuclear construction there, showing that nuclear energy would be the least-cost option for new generating capacity. The study compared nuclear, coal, gas turbine combined cycle and peat. Nuclear has very much higher capital costs than the others – EUR 1749/kW including initial fuel load, which is about three times the cost of the gas plant. But its fuel costs are much lower, and so at capacity factors above 64% it is the cheapest option.

August 2003 figures put nuclear costs at EUR 2.37 c/kWh, coal 2.81 c/kWh and natural gas at 3.23 c/kWh (on the basis of 91% capacity factor, 5% interest rate, 40 year plant life). With emission trading @ EUR 20/t CO2, the electricity prices for coal and gas increase to 4.43 and 3.92 c/kWh respectively:

 Projected Electricity Costs for Finland 2003

In the middle three bars of this graph the relative effects of capital and fuel costs can be clearly seen. The relatively high capital cost of nuclear power means that financing cost and time taken in construction are critical, relative to gas and even coal. But the fuel cost is very much lower, and so once a plant is built its cost of production is very much more predictable than for gas or even coal. The impact of adding a cost or carbon emissions can also be seen.

In 2013 the Nuclear Energy Institute announced the results of its financial modelling of comparative costs in the USA, based on figures from the US Energy Information Administration’s 2013 Annual Energy Outlook. NEI assumed 5% cost of debt, 15% return on equity and a 70/30 debt equity capital structure. The figures are tabulated below. The report went on to show that with nuclear plant licence renewal beyond 60 years, power costs would be $53-60/MWh.

NEI 2013 Financial Modelling

  EPC cost capacity Electricity cost
Gas combined cycle, gas @ $3.70/GJ $1000/kW 90% $44.00/MWh
Gas combined cycle, gas @ $5.28/GJ $1000/kW 90% $54.70/MWh
Gas combined cycle, gas @ $6.70/GJ $1000/kW 90% $61.70/MWh
Gas combined cycle, gas @ $6.70/GJ, 50-50 debt-equity $1000/kW 90% c $70/MWh
Supercritical pulverised coal, 1300 MWe $3000/kW 85% $75.70/MWh
Integrated gasification combined cycle coal, 1200 MWe $3800/kW 85% $94.30/MWh
Nuclear, 1400 MWe (EIA's EPC figure) $5500/kW 90% $121.90/MWh
Nuclear, 1400 MWe (NEI suggested EPC figure) $4500-5000/kW 90% $85-90/MWh
Wind farm, 100 MWe $1000/kW 30% 112.90/MWh

5% cost of debt, 15% return on equity and a 70-30 debt equity capital structure.

The China Nuclear Energy Association estimated in May 2013 that the construction cost for two AP1000 units at Sanmen are CNY 40.1 billion ($6.54 billion), or 16,000 Yuan/kW installed ($2615/kW) – about 20% higher than that of improved Generation II Chinese reactors, but likely to drop to about CNY 13,000/kW ($2120/kW) with series construction and localisation as envisaged. Grid purchase price is expected to exceed CNY 0.45/kWh at present costs, and drop to 0.42 with reduced capital cost.

A striking indication of the impact of financing costs is given by Georgia Power, which said in mid 2008 that twin 1100 MWe AP1000 reactors would cost $9.6 billion if they could be financed progressively by ratepayers, or $14 billion if not. This gives $4363 or $6360 per kilowatt including all other owners costs.

Finally, in the USA the question of whether a project is subject to regulated cost recovery or is a merchant plant is relevant, since it introduces political, financial and tactical factors. If the new build cost escalates (or is inflated), some cost recovery may be possible through higher rates can be charged by the utility if those costs are deemed prudent by the relevant regulator. By way of contrast, a merchant plant has to sell all its power competitively, so must convince its shareholders that it has a good economic case for moving forward with a new nuclear unit.

Providing investment incentives

As more electricity markets become deregulated and competitive, balancing supply and demand over the short-term can result in significant price volatility. Price signals in the spot market for electricity supply do not provide a guide on the return that might be achieved over the long term, and fail to create incentive for long-term investment in generation or transmission infrastructure, nor do they value diversity of supply.

Deregulated electricity markets with preferential grid access for renewables have left some utilities with stranded assets, which can no longer be used sufficiently fully to be profitable. As a result, many are being decommissioned, eg about 9 GWe by E.On and RWE in Germany to 2013, and a further 7.3 GWe expected there (apart from nuclear capacity).

The economic rationale for electricity from any plants with high capital cost and long life does not translate into incentive for investment unless some long-term electricity price is assured. This has been tackled differently in various countries.

In the USA, investment in new capital-intensive plant is going ahead only in states where cost-recovery can be assured, not in deregulated areas. Proposed merchant plants in Texas and some eastern states have been postponed indefinitely.

In Ontario, Canada, the refurbishment of Bruce A 1&2 was underwritten by a power purchase agreement (PPA) at about $63/MWh, slightly higher than the regulated price. The refurbishment of Bruce A 3&4 (1500 MWe) from 2016 and the C$ 8 billion needed for the Bruce B units (3480 MWe) from 2020 is likely to be underwritten similarly with PPAs.

In the UK, legislation from 2013 has three main elements:

  • Feed-in tariffs (FIT), now relatively common in several countries, give particular low-carbon producers a predictable return per kWh over a set period regardless of prevailing market prices. The FIT can take several forms. In the UK it will be effected through contracts for difference (CfD) which remove long-term exposure to electricity price volatility. The FIT with CfD means that if the market price is lower that the agreed ‘strike price’, the government or the transmission system operator (TSO) pays that difference per kWh, if the market is above the strike price the generator pays the TSO or government. They are long-term contracts which can be capped regarding quantity of power, helping developers secure the large upfront capital costs for low-carbon infrastructure while protecting consumers from rising energy bills. The first strike prices were published in the 2013-18 Delivery Plan: £155/MWh for offshore wind, £100/MWh for onshore wind and £125/MWh for large solar PV.
  • A floor price for ‘carbon’ to support de-carbonisation. The idea is that a carbon floor price will drive the market towards any FIT or strike price level applied to clean sources.
  • Capacity market measures will be introduced. These involve payments for dispatchable capacity maintained to ensure that demand can be met regardless of short-term conditions affecting other generators. They will work through penalties and availability payments to provide incentive for generators to be available when needed, in effect paying for reliability. The first capacity auction is scheduled for late 2014, for delivery during winter 2018-19.

In October 2013 the UK government announced that initial agreement had been reached with EDF Group on the key terms of a proposed £16 billion investment contract for the Hinkley Point C nuclear power station. The key terms include 35-year ‘Contract for Difference’, the Strike Price of £89.50/MWh being fully indexed to the Consumer Price Index and conditional upon the Sizewell C project proceeding. If it does not for any reason, and the developer cannot share first-of-a-kind costs across both, the strike price is to be £92.50/MWh. EDF said that the agreement in principle is not legally binding, and is dependent on a positive decision from the European Commission in relation to State Aid, following which it will make a final decision on the project. As well as price per MWh, the question of guaranteed load factor arises so that output is sufficient to amortise the investment, in the face of renewables’ preferential grid access.

In the Czech Republic, CEZ says that investment in two new 1200 MWe reactors at Temelin will not proceed unless it has some assurance of long-term electricity prices from them. The government and Ministry of Finance are resisting this. The Industry Ministry was proposing €60/MWh, others suggest that €90 will be needed, indexed. CEZ requires €70/MWh for the new units to be profitable, compared with mid-2013 forward prices of under €40.

In Turkey, in order to secure investment in the 4x1200 MWe Akkuyu nuclear power plant, a formula for long-term power prices was worked out. This involves the Turkish Electricity Trade & Contract Corporation (TETAS) buying a fixed proportion of the power at a fixed price of US$ 123.50/MWh for 15 years, or to 2030. The proportion will be 70% of the output of the first two units and 30% of that from units 3&4 over 15 years from commercial operation of each. Rosatom will initially have full ownership of the project.

Sources:

OECD/ IEA NEA 2010, Projected Costs of Generating Electricity.
OECD, 1994, The Economics of the Nuclear Fuel Cycle.
OECD/NEA, 2012, Nuclear Energy and Renewables - System Effects in Low-carbon Electricity Systems.
US Energy Information Administration, 2013, Levelized Cost of New Generation Resources in the Annual Energy Outlook 2013.
NEI: US generating cost data.
Tarjanne, R & Rissanen, S, 2000, Nuclear Power: Least-cost option for base-load electricity in Finland; in Proceedings 25th International Symposium, Uranium Institute.
Gutierrez, J 2003, Nuclear Fuel – key for the competitiveness of nuclear energy in Spain, WNA Symposium.
University of Chicago, August 2004, The Economic Future of Nuclear Power.
Nuclear Energy Institute, August 2008, The cost of new generating capacity in perspective.
Richard Myers, Feb 2013, Nuclear Energy in 2013: Status and Outlook, Remarks to NEI’s License Renewal Workshop
Direction Générale de l'Energie et du Climat, 2008, Synthèse publique de l'étude des coûts de référence de la production électrique
ExternE website
Xu Yuming, May 2013, China’s Nuclear Power Development in Post-Fukushima Era, CNEA.
Ron Cameron, OECD/NEA, July 2013 presentations to Australian Academy of Technological Sciences & Engineering conference in Sydney.
Nuclear Energy Institute presentation to financial markets, February 2014
Nadira Barkatullah, Financing Nuclear Power Projects: Challenges and Approaches, World Nuclear Association 2014 Symposium

Appendix

Some other figures in recent years

There have been a large number of recent estimates from the United States of the costs of new nuclear power plants. For example, Florida Power & Light in February 2008 released projected figures for two new AP1000 reactors at its proposed Turkey Point site. These took into account increases of some 50% in material, equipment and labour since 2004. The new figures for overnight capital cost ranged from $2444 to $3582 /kW, or when grossed up to include cooling towers, site works, land costs, transmission costs and risk management, the total cost came to $3108 to $4540 per kilowatt. Adding in finance charges almost doubled the overall figures at $5780 to $8071 /kW. FPL said that alternatives to nuclear for the plant were not economically attractive. 

In May 2008 South Carolina Electric and Gas Co. and Santee Cooper locked in the price and schedule of new reactors for their Summer plant in South Carolina at $9.8 billion. (The budgeted cost earlier in the process was $10.8 billion, but some construction and material costs ended up less than projected.) The EPC contract for completing two 1,117-MW AP1000s is with Westinghouse and the Shaw Group. Beyond the cost of the actual plants, the figure includes forecast inflation and owners' costs for site preparation, contingencies and project financing. The units are expected to be in commercial operation in 2016 and 2019.

In November 2008 Duke Energy Carolinas raised the cost estimate for its Lee plant (2 x 1117 MWe AP1000) to $11 billion, excluding finance and inflation, but apparently including other owners costs.

In November 2008 TVA updated its estimates for Bellefonte units 3 & 4 for which it had submitted a COL application for twin AP1000 reactors, total 2234 MWe. It said that overnight capital cost estimates ranged from $2516 to $4649/kW for a combined construction cost of $5.6 to 10.4 billion. Total cost to the owners would be $9.9 to $17.5 billion.

Regarding bare plant costs, some recent figures apparently for overnight capital cost (or Engineering, Procurement and Construction – EPC – cost) quoted from reputable sources but not necessarily comparable are:

EdF Flamanville EPR: EUR 4 billion/$5.6 billion, so EUR 2434/kW or $3400/kW
Bruce Power Alberta 2x1100 MWe ACR, $6.2 billion, so $2800/kW
CGNPC Hongyanhe 4x1080 CPR-1000 $6.6 billion, so $1530/kW
AEO Novovronezh 6&7 2136 MWe net for $5 billion, so $2340/kW
AEP Volgodonsk 3 & 4, 2 x 1200 MWe VVER $4.8 billion, so $2000/kW 
KHNP Shin Kori 3&4 1350 MWe APR-1400 for $5 billion, so $1850/kW
FPL Turkey Point 2 x 1100 MWe AP1000 $2444 to $3582/kW
Progress Energy Levy county 2 x 1105 MWe AP1000 $3462/kW
NRG South Texas 2 x 1350 MWe ABWR $8 billion, so $2900/kW
ENEC for UAE from Kepco, 4 x 1400 MWe APR-1400 $20.4 billion, so $3643/kW